Phosphabenzenes as electron withdrawing phosphine ligands in catalysis

Article abstract of DOI:10.1039/b101454o

The utility of phosphabenzenes as ligands in late transition metal catalysis is examined. Molecular orbital calculations indicate that phosphabenzenes possess a low lying LUMO permitting π-back bonding interactions. The resultant electron withdrawing nature of the phosphabenzenes is beneficial for reactions in which reductive elimination steps are rate-limiting. For example, phosphabenzenes were found to function well in nickel catalyzed [4 + 2] cycloadditions.

Full text of DOI:10.1039/b101454o

View Article Online / Journal Homepage / Table of Contents for this issue
Phosphabenzenes as electron withdrawing phosphine ligands in
catalysis
Erin F. DiMauro and Marisa C. Kozlowski*
1
Department of Chemistry, Roy and Diana Vagelos Laboratories, University of Pennsylvania,
Philadelphia, Pennsylvania 19104, USA
Received (in Cambridge, UK) 13th February 2001, Accepted 28th November 2001
First published as an Advance Article on the web 21st December 2001
The utility of phosphabenzenes as ligands in late transition metal catalysis is examined. Molecular orbital
calculations indicate that phosphabenzenes possess a low lying LUMO permitting π-back bonding interactions.
The resultant electron withdrawing nature of the phosphabenzenes is beneficial for reactions in which reductive
elimination steps are rate-limiting. For example, phosphabenzenes were found to function well in nickel catalyzed
[4 ϩ 2] cycloadditions.
Introduction
Our interest in phosphabenzene derivatives (Fig. 1) as potential
Fig. 1 Binding modes of phosphabenzene.
ligand candidates for transition metal catalyzed reactions was
stimulated by several factors, including the novelty of such
applications, the ease of preparation of ligand candidates, their
stability, and the unique electronic properties of these systems.
Many complexes of phosphabenzenes with transition metals
are known.1 The coordination mode of phosphabenzenes may
vary according to the metal to which they are bound. In general
there are three conventional binding modes: η1, η6, and mixed
(Fig. 1). The η1 binding mode is typically observed with late
transition metals in low oxidation states2 while η6 is favored by
early transition metals in high oxidation states.3 Metals in the
center of the transition series tend to exhibit mixed binding.4
Despite the prevalence of reported transition metal–
phosphabenzene complexes, phosphabenzenes remain relatively
unexplored as potential ligands in catalytic cycles. One impor-
tant advancement in this area has been made by Breit et al. who
demonstrated the utility of phosphabenzenes as ligands in the
Rh-catalyzed hydroformylation of styrene. In this reaction
2,4,6-trisubstituted phosphabenzenes provided high activity
and excellent regioselectivity for the branched aldehyde.5
Fig.
2
B3LYP/6-31ϩG** molecular orbitals of phosphabenzene
(orbital energies in eV are shown parenthetically).
would participate in η6-coordination. The fact that these
orbitals are close in energy undoubtedly accounts for the range
of coordination modes observed in various phosphabenzene
metal complexes. The HOMO2Ϫ donor orbital is complemented
by the LUMO which enables phosphabenzene to act as a π-
acceptor ligand when coordinated to a metal center through
phosphorus. Based on these orbitals, phosphabenzenes in η1-
metal complexes are expected to act as electron withdrawing
ligands which can stabilize metals in lower oxidation states.
Tolman has introduced the parameters χ and θ to quantify
the electronic and steric properties of ligands.11 A large value of
χ indicates that the ligand is a strong π-acceptor, while a small
value of χ is indicative of strong σ-donation. The parameter θ
provides a direct measure of ligand cone angle which is used as
a measure of the volume around the metal occupied by the
ligand.
Results and discussion
Phosphabenzene steric and electronic properties
A number of theoretical studies have been reported for phos-
phabenzenes6,7 which indicate that both high energy occupied
and low energy unoccupied molecular orbitals have large co-
efficients at phosphorus.8 From B3LYP/6-31ϩG** calculations
of the parent phosphabenzene,9 we generated the schematics10
of these orbitals as illustrated in Fig. 2. These representations
indicate that phosphabenzene possesses a high lying orbital
(HOMO2Ϫ) suitable for σ-metal coordination which is close in
energy to the π-bonding HOMO1Ϫ and HOMO orbitals that
In order to assess the steric properties of the phospha-
benzenes, we calculated values of θ for 1–4 following a protocol
similar to that employed by Tolman for unsymmetric tri-
substituted phosphines. In this study, θ values were obtained
by averaging the phosphabenzene occupancy angles (α and β)
along two orthogonal planes (x and y) which are shown in
DOI: 10.1039/b101454o
J. Chem. Soc., Perkin Trans. 1, 2002, 439–444
This journal is © The Royal Society of Chemistry 2002
439

View Article Online
Table 1 Carbonyl IR shifts of Cr(CO)5 complexes of phosphaben-
zenes (Eqn. (2))
IR/cmϪ1 pentanea, b
IR /cmϪ1 CHCl3
a, c
Ligand
1
2066.9
1994.1
1937.0
2066.7
1993.8
1933.2
2066.0
1992.5
1929.1
—
2066.9 (2068.1)
1995.7
1937.7
2066.8 (2068.1)
1994.3
1932.1
2065.8 (2067.6)
1995.4
1926.8
2
3
4d
—
a Cast as a thin film on NaCl plates unless otherwise noted. b Mother
liquor from pentane crystallization. c Crystals from pentane redissolved
in CHCl3. Parenthetic entries are measurements of CHCl3 solution.
d Complex did not form.
Fig. 3
Table 2 Calculated Mulliken and natural bond order charges for the
phosphorus in 1–4
Fig. 3. The occupancy angles, α and β, were measured from the
geometries of the calculated structures.12 In the conformations
of 1–4 used in this analysis, the 2,6-substituents were oriented
to minimize steric interactions with respect to phosphabenzene
ring and a σ-bound metal; representative conformations used
for 1 and 4 are illustrated in Fig. 3. While the steric demand
is relatively small in the y plane (small β), the demand is very
large in the x plane (large α). As such, the space occupied by
phosphabenzenes is better characterized as a flattened cone
rather than a symmetric cone. While the average parameter θ
(see Table 3 below) gives a general idea as to the steric demand
of these compounds, consideration of α and/or β may be more
appropriate depending on the coordination geometry of the
metal and the rotational mobility about the phosphorus–metal
bond. The latter may be partially fixed by π-backbonding inter-
actions. In addition, the geometric arrangements of more
than one phosphabenzene around a metal center are limited
such that the steric factor θ may be underestimated in this
analysis.
HF/6-31 G*b
Mulliken
B3LYP/6-31G*b
PM3a
Mulliken
Ligand
NBO
0.71
Mulliken
0.21
NBO
0.47
0.31
0.66
1
2
3
4
0.45
0.47
0.50
0.42
0.31
0.29
0.26
0.26
0.72
0.69
0.67
0.66
0.19
0.17
0.15
0.13
0.66
0.65
0.64
0.62
a Calculated using SPARTAN. Mulliken and NBO charges were identi-
cal. b Calculated using Gaussian 98.
charges listed in Table 2.15,16 The B3LYP/6-31G* Mulliken
charges obtained in this study for the parent phosphabenzene 1
are the same as those previously reported.17 The trends between
the different derivatives are the same regardless of the method
used; phosphabenzenes containing alkyl groups tend to have
smaller positive charges at phosphorus due to the donating
nature of these groups. Overall, the charges are fairly similar
between the different derivatives which should translate to
similar χ values for these ligands (see Table 3 below). The
phosphabenzenes with more alkyl groups (3, 4) have slightly
greater electron density consistent with their slightly smaller χ
values (more donating).
The χ parameter for 3 was estimated by Breit et al.5a based
upon the infrared CO stretching frequencies in the η1 Cr(CO)5
complex with 113,14 (Eqn (1), and Eqn. (2)). In order to deter-
(1)
The χ and θ values measured in this study for 1–4 are
collected in Table 3 together with the values for some
common phosphorus ligands. From these values the phospha-
benzenes are expected to act as moderately electron withdraw-
ing ligands with properties similar to phosphites. In addition,
the steric size of these compounds can be adjusted by changing
the 2,6-substituents without substantially altering the electronic
properties.
(2)
mine if this electronic parameter varies significantly upon
substitution of the phosphabenzene ring with aryl or alkyl
groups, we measured the stretching frequencies for the Cr(CO)5
complexes of 1–4 (Table 1). For the hindered phosphabenzene
4, the Cr(CO)5 complex did not form indicating that the
two tert-butyl groups ortho to the phosphorus sterically inhibit
coordination. From measurements of the remaining com-
plexes, it appears that the presence of aryl vs. alkyl groups on
the phosphabenzene ring has a relatively small influence on
the electron-withdrawing character of the phosphorus,
and 1–3 are expected to have similar χ values between 23–24
which are comparable to those of phosphites (see Table 3
below).
Phosphabenzenes in nickel-catalyzed cycloisomerization
Based on the electronic properties described above, phospha-
benzenes are anticipated to affect the rate of a catalytic cycle
through a strong electronic influence at the metal center. There-
fore, a catalytic system was examined in which the rate deter-
mining step involves an increase in electron density at the metal
center. The phosphabenzene ligands should stabilize the inter-
mediates at this step and enhance the reaction. The formal
[4 ϩ 2] cycloadditions of dieneynes developed by Wender18 had
the necessary prerequisites since the rate determining step is
most likely reductive elimination of the product from the metal
center (Scheme 1). The reported tetrahedral complex involving
These χ value estimates are consistent with charge calcula-
tions for 1–4 in which the phosphorus atoms have the partial
440
J. Chem. Soc., Perkin Trans. 1, 2002, 439–444

View Article Online
Table 3 Steric and electronic parameters for representative phos-
phorus ligands
Entry
Ligand
χa
θ/dega
1
2
3
4
5
6
7
8
9
PPh3
PEt3
PCy3
P(OMe)3
P(OPh)3
P(O-o-C6H4Ph)3
P(O-i-C3HF6)3
1
2
3
4
12.9
5.6
0.3
145
132
170
107
128
152
130
135e
123e
117e
128e
23.3
29.1
28.9
51.3
24b
24c
10
11
23c
∼23c, d
Scheme 2 Synthesis of phosphabenzenes.
a Data for entries 1–7 from Tolman et al. (Ref. 11). χ values obtained
using the IR stretches of the Ni(CO)4 complexes. b See Ref. 5a. c Esti-
mated from LؒCr(CO)5 IR stretches in this work (see text). d Since
4ؒCr(CO)5 did not form, χ could not be measured directly for 4. The
indicated value was estimated from 3. e Determined from calculated
structures in this work (see text).
The results of the phosphabenzene ligands in the nickel cata-
lyzed cycloaddition (Eqn. 3) are shown in Table 4. Background
(3)
thermal reactions in both THF and cyclohexane (entries 1 and
2) indicated that the thermal process was slow and caused
decomposition as well. All of the catalytic reactions afforded
the product as a single diastereomer. As part of a control, the
more hindered P(O-o-C6H4Ph)3 phosphite ligand was found
to be superior to P(OPh)3 (entry 3 vs. 4). Use of 2,4,6-tri-
phenylphosphabenzene (1) gave unexpectedly poor conversion
(entry 8, 7% after 20 h) when the optimal conditions for the
phosphite ligands were employed (30 mol% ligand, THF).
Reasoning that the coordinating THF solvent may be inter-
fering with coordination of the highly hindered 1, the reaction
was also examined in cyclohexane. Gratifyingly, this change
resulted in much faster conversions (entry 9, 63% after 20 h).
Upon close analysis of the catalytic cycle (Scheme 1), only
two phosphine ligands appear to be needed for coordination
to nickel. Examination of the 1 : Ni(COD)2 ratio revealed that a
2 : 1 ratio provided much better rates (92% after 20 h, entry 10)
than a 3 : 1 or 1 : 1 ratio (63% and 25% conversion respectively;
entries 9 and 12).
Scheme 1 Mechanism of the nickel catalyzed [4 ϩ 2] cycloaddition.
four phosphabenzene ligands bound to Ni(0) in an η1-mode,
indicates that phosphabenzenes are suitable modifiers for the
Ni(0) catalysts employed in this reaction.19
Since the use of the minimal amount of ligand and a non-
coordinating solvent was found to be effective with the phos-
phabenzene ligands, these parameters were reexamined for
the P(O-o-C6H4Ph)3 case. Interestingly, the use of the non-
coordinating solvent cyclohexane in place of THF led to poorer
results (entry 6 vs. 4) in contrast to the observation for phos-
phabenzene 1. To distinguish between a bulk solvent effect and
a potential ligand role for the THF, the reaction was performed
with the addition of 50 mol% THF to the conditions in entry 6
(entry 7). The reaction rate under these conditions was the same
as when THF was the solvent, which further supports the
hypothesis that THF coordinates to the nickel. In line with this
hypothesis, an excess of P(O-o-C6H4Ph)3 may be required to
ensure a favorable equilibrium toward a 2 : 1 ligand : Ni com-
plex when THF is present. Indeed, the minimal amount of
phosphite (2 : 1 ligand : Ni) was less effective (entry 4 vs. 5) than
when an excess was used (3 : 1 ligand : Ni), which again is the
opposite of what was observed with 1. Notably, the rates
achieved under the optimal conditions with phosphabenzene 1
(entry 10) are superior to those for P(OPh)3 and comparable to
those for the more sterically congested P(O-o-C6H4Ph)3. Since
the phosphabenzene ligands are stable toward air and even
chromatography, the use of such ligands may be advantageous
in situations where the less stable phosphite ligands cannot be
tolerated.
The phosphabenzenes examined in this study were chosen on
the basis of ease of preparation, stability and symmetry. The
series, 1–4, with phenyl, methyl, and tert-butyl substituents
permits a study of substituent steric and electronic effects on
reactivity. In the series 1 to 3 the cone angle decreases as the
ortho-phenyl groups are replaced with methyl groups while 4
(ortho tert-butyl groups) has a similar cone angle (but different
electronic properties) compared to 1. Although several methods
have been reported for synthesis of phosphabenzenes,20 the
unified protocols presented in Scheme 2 were found to be the
most convenient for preparing these compounds. For phos-
phabenzenes 1–3, the general method Märkl et al.21 reported for
1 was simplest, in which the pyrylium iodide salts are converted
to the phosphabenzenes upon treatment with tris(trimethyl-
silyl)phosphine. We found that exchange of the pyrylium
tetrafluoroborates by treatment with KI in HOAc to the
required pyrylium iodide salts was more convenient than direct
preparation of the iodides.22 The tetrafluoroborate precursors
7
23 and 824 were synthesized following reported protocols while
we prepared 9 by adapting a procedure used previously for the
corresponding perchlorate salt.25 For the tris(tert-butyl)-
phosphabenzene 4, these methods could not be employed and
the alternate procedure with tris(hydroxymethyl)phosphine26 as
reported by Dimroth et al.27 was used.
J. Chem. Soc., Perkin Trans. 1, 2002, 439–444
441

View Article Online
Table 4 Conversion vs. phosphine in the nickel-catalyzed isomerization (Eqn. 3)
Entry
Mol% liganda
Ligand
Solventb
% conversionc
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
—
—
30
30
20
30
30
30
30
20
20
10
20
20
10
20
30
—
—
THF
Cyclohexane
THF
THF
THF
Cyclohexane
Cyclohexanef
THF
Cyclohexane
Cyclohexane
Cyclohexaneg
Cyclohexane
Cyclohexane
Cyclohexane
Cyclohexane
Cyclohexane
Cyclohexane
∼4d
10d, e
18
P(OPh)3
P(O-o-C6H4Ph)3
P(O-o-C6H4Ph)3
P(O-o-C6H4Ph)3
88
∼40d
70
P(O-o-C6H4Ph)3
89
7
63
92
89
25
81
66
1
1
1
1
1
2
3
4
4
4
19
31
22
a 10 mol% Ni(COD)2 used except for entries 1 and 2. b Reactions in THF were at 66 ЊC and in cyclohexane at 80 ЊC unless otherwise noted.
c Conversions determined after 20 h by 1H NMR. No substrate decomposition or by products observed unless indicated. d Significant decomposition
observed. As such, the reported conversions represent an upper limit. e Isolated yield. 16% conversion by 1H NMR. f 50 mol% THF. g Reaction run at
66 ЊC.
The phenyl groups (R1, R3) adjacent to the phosphorus in the
phosphabenzenes play a key role. With derivative 2, where one
of these phenyl groups is replaced by Me, the reaction still pro-
ceeds well (entry 13). When only alkyl groups are present as in 3
(R1, R3 = Me) or 4 (R1, R3 = t-Bu), poor rates are observed
(entries 14 and 16). Comparison of the cone angles to the yields
for 1 (135Њ, 92%), 2 (123Њ, 81%), 3 (117Њ, 66%), and 4 (128Њ,
31%) indicates that the composition of the adjacent group
(phenyl vs. alkyl) is as important as the steric size of the ligand.
At least one ortho-phenyl group is needed to ensure high
reactivity. This result is at first puzzling, since the orthogonal
phenyl groups cannot engage in electron withdrawal through
conjugative interaction with the phosphabenzene ring. How-
ever, the sp2-hybridized carbons of the ortho-phenyl groups in 1
and 2 are electron withdrawing compared to the sp3-hybridized
carbons of the ortho-alkyl groups in 2, 3, and 4 which may be
influencing the electronic character of the phosphabenzene
ring.
oven (90 ЊC) for at least six hours. When necessary, solvents and
reagents were dried prior to use. Toluene, Et2O, THF, and
CH2Cl2 were de-oxygenated by purging with Ar and then dried
by passing through activated alumina. Cyclohexane and THF
for the Ni(0)-catalyzed reactions were distilled from Na–
benzophenone ketyl and were further degassed in the reaction
flask by bubbling N2 through the solution. TMSCl, iPr2NH,
CH3CN, pyridine, and HMPA were distilled from CaH2.
Anisole was distilled from Na. NaH was rinsed three times with
hexanes (distilled from CaH2) and dried under high vacuum
prior to use. Ni(COD)2 was purchased from Strem and was
stored under an N2 atmosphere in a dry box. Tris(trimethyl-
silyl)phosphine was purchased from Aldrich. Phosphabenzenes
1
21 and 427 were prepared following the published procedures.
Analytical thin layer chromatography (TLC) was performed
on EM reagents 0.25 mm silica-gel 60-F plates. Visualization
was accomplished with UV light. Chromatography on silica gel
was performed using a forced flow of the indicated solvent
system on EM reagents silica gel 60 (230–400 mesh).28
1H NMR spectra were recorded on Bruker AM-500 (500
MHz), AM-250 (250 MHz), or AM-200 (200 MHz) spectro-
meters. Chemical shifts are reported in ppm from tetramethyl-
silane (0 ppm) or with the solvent resonance as the internal
standard (CDCl3 7.26 ppm, DMSO-d6 2.49 ppm, D2O 4.80
ppm). Data are reported as follows: chemical shift, multiplicity
(s = singlet, d = doublet, t = triplet, q = quartet, br = broad,
m = multiplet), coupling constants, and number of protons.
Mass spectra were obtained on a low resolution Micromass
Platform LC in electron spray mode. IR spectra were taken on a
Perkin–Elmer FT-IR spectrometer using thin films or in a cell
using CHCl3 solution with CHCl3 as background. Melting
points were obtained on Thomas Scientific Unimelt apparatus
and are uncorrected.
Conclusion
The results obtained thus far indicate that the rate of this reac-
tion is influenced by both the steric and electronic properties of
the phosphabenzene ligands. As expected, the reaction is
accelerated by a moderate increase in the steric demand of the
ligand and a decrease in the electron density at nickel. Tuning
of the electronic and steric properties of phosphabenzene can
be accomplished by simple structural modifications allowing
optimal combinations of χ and θ to be identified. While the
exact role of solvent in the catalytic cycle of the nickel catalyzed
cyclizations remains unclear, a significant rate enhancement is
seen with phosphabenzene upon switching from a coordinating
solvent to a non-coordinating solvent. This solvent effect is
contrary to that observed for the phosphite cases, in which the
maximum rate is obtained when another coordinating species is
present. These results indicate that phosphabenzenes would be
useful ligands in applications which call for bulky, electron-
deficient ligands that are stable to air and water.
2,6-Dimethyl-4-phenylpyrylium tetrafluoroborate (9)
Compound 9 was synthesized by modification of a reported
procedure for a related pyrylium.29 2,6-Dimethyl-4-pyrone
(5.0 g, 0.04 mol) was dissolved in anisole (40 mL) with gentle
heating. The solution was cooled to 5 ЊC and PhMgBr
(0.04 mol) was added slowly as a 1 M solution in ether via
cannula. The purple solution was then warmed to room
temperature and poured over HBF4 (6.7 mL of a 54% solution
in ether, 0.048 mol). The red–orange precipitate was isolated
by filtration and washed thoroughly with hot EtOH to yield 9
as a pink–orange solid (6.5 g) in 59% yield.
Experimental
General
Unless otherwise noted, all non-aqueous reactions were carried
out under an atmosphere of dry nitrogen in glassware that had
been either flame dried in a stream of nitrogen or dried in an
442
J. Chem. Soc., Perkin Trans. 1, 2002, 439–444

View Article Online
Synthesis of pyrylium iodides from pyrylium tetrafluoroborates
7–9
1H NMR (CHCl3, 500 MHz) δ 5.65 (s, 2H), 4.4 (d, J = 11.7 Hz,
1H), 4.2 (d, J = 11.7 Hz, 1H), 4.1 (t, J = 7.8 Hz, 1H), 3.25
(dd, J = 5, 11.3 Hz, 1H), 2.9 (m, 1H), 1.19 (d, J = 7.2 Hz, 3H),
0.17 (s, 9H).
To a warm (40–50 ЊC) slurry of the pyrylium tetrafluoroborate
(3.8 mmol) in acidified (2–4 drops HOAc) water (20 mL)
was added KI (3.8 mmol). The solution was stirred until red
(1–5 h) then cooled to room temperature. The iodide was
collected by filtration and washed thoroughly with ether. Yields
are quantitative.
General procedure for Ni(0)-catalyzed carbocyclization
Dieneyne 5 (60 mg, 0.28 mmol) and 1 (28 mg, 0.086 mmol) were
combined in an acid washed, base washed, flame dried Schlenk
flask in an inert atmosphere glovebox and dissolved in distilled
THF or cyclohexane (25 mL). A solution of Ni(COD)2 (7.9 mg,
0.028 mmol) in distilled THF or cyclohexane (4 mL) was then
added to the stirred solution, dropwise. The reaction flask was
then removed from the glovebox and heated to reflux. Since
there was no decomposition, the reaction progress could be
6-Methyl-2,4-diphenylphosphabenzene (2) and 2,6-dimethyl-4-
phenyl phosphabenzene (3)
Following the method that Märkl et al.21 employed for the
synthesis of 1, a solution of the pyrylium iodide precursor
(2.5 mmol) in dry CH3CN (15 mL) was purged with N2. P-
(TMS)3 (2.8 mmol) was added via gas-tight syringe and the
solution was heated at reflux for 20 h. After cooling, CH3CN
was removed in vacuo and the crude product was purified by
flash chromatography on SiO2 (0–5% EtOAc–hexanes) to yield
the phosphabenzenes 2 and 3 as pale yellow powders in 46%
yield and 67% yield respectively. 2: mp 79–80 ЊC; 1H NMR (500
MHz, CDCl3) δ 8.1 (d, J = 5.4 Hz, 1H), 7.9 (d, J = 7.0 Hz, 1H),
7.7 (d, J = 8.2 Hz, 2H), 7.6 (d, J = 9.3 Hz, 2H), 7.35 (m, 6H),
2.8 (d, J = 15.4 Hz, 3H); 13C NMR (125 MHz, CDCl3) δ 172.0
(d, J = 50 Hz), 169.0 (d, J = 50 Hz), 144.1 (d, J = 5 Hz), 144.0,
143.8, 133.4 (d, J = 12.5 Hz), 131.3 (d, J = 12.5 Hz), 129.3 (d,
J = 10 Hz), 128.1, 127.9, 25.2 (d, J = 37.5 Hz); 31P NMR (250
1
monitored by H NMR. Conversion was determined by com-
paring the relative integration for the peaks at δ 1.8 (d, CH3, 5)
and δ 1.19 (d, CH3, 6).
Acknowledgements
Financial support was provided by the University of Pennsyl-
vania, the National Science Foundation (CHE-9730576),
Merck Research Laboratories, and DuPont. Acknowledgment
is made to the donors of the Petroleum Research Fund,
administered by the American Chemical Society, for partial
support of this research. We thank Dr Manoranjan Panda for
carrying out the calculations.
1
MHz, CDCl3) δ 188.76. 3: mp 49–50 ЊC; H NMR (500 MHz,
CDCl3) δ 7.75 (d, J = 6.7 Hz, 2H), 7.62 (d, J = 7.9 Hz, 2H), 7.46
(t, J = 7.0 Hz, 2H), 7.4 (t, J = 6.7 Hz, 1H), 2.73 (d, J = 15.1 Hz,
6H); 13C NMR (125 MHz, CDCl3) δ 168.8 (d, J = 49 Hz), 143.8,
142.8, 142.7, 132.5 (d, J = 13.7 Hz), 129.2, 128.04, 25.0 (d,
J = 36.2 Hz); 31P NMR (250 MHz, CDCl3) δ 193.3. Phospha-
benzene 3 has previously been synthesized by a different
method.30
References
1 (a) P. Le Floch and F. Mathey, Coord. Chem. Rev., 1998, 179–180,
771–791; (b) N. Mezailles, F. Mathey and P. Le Floch, Prog. Inorg.
Chem., 2001, 49, 455–550.
2 (a) H. Kanter and K. Dimroth, Tetrahedron Lett., 1975, 541–544;
(b) F. Mathey and P. Le Floch, Chem. Ber., 1996, 263–268;
(c) C. Elschenbroich, S. Voss and K. Harms, Z. Naturforsch, 1999,
54b, 209–213.
1-(3-Trimethylsilylprop-2-ynyloxy)hexa-2,4-diene (5)
3 (a) F. Knoch, F. Kremer, U. Schmidt, U. Zenneck, P. Le Floch and
F. Mathey, Organometallics, 1996, 15, 2713–2719; (b) P. Le Floch,
F. Knoch, F. Kremer, F. Mathey, W. Scholtz, K.-H. Thiele and
U. Zenneck, Eur. J. Inorg. Chem., 1998, 119–126; (c) C. Elschen-
broich, M. Nowotony, B. Metz, W. Massa, J. Graulich, K. Biehler
and W. Sauer, Angew. Chem., Int. Ed., 1991, 30, 547–550.
4 (a) F. Nief, C. Charrier, F. Mathey and M. Simalty, J. Organomet.
Chem., 1980, 187, 277–285; (b) K. C. Nainan and C. T. Sears,
J. Organomet. Chem., 1978, 148, C31–C34.
5 (a) B. Breit, Chem. Commun., 1996, 2071–2072; (b) B. Breit,
R. Winde and K. Harms, J. Chem. Soc., Perkin Trans. 1, 1997,
2681–2682; (c) B. Breit, J. Mol. Catal. A: Chem., 1999, 143–154;
(d ) B. Breit, R. Winde, T. Mackewitz, R. Paciello and K. Harms,
Chem. Eur. J., 2001, 7, 3106–3121.
6 For a general review A. J. Ashe III, Acc. Chem. Res., 1978, 11,
153–157.
7 (a) V. Jonas and G. Frenking, Chem. Phys. Lett., 1993, 210, 211–215;
(b) J. Lorentzon, M. P. Fülscher and B. O. Roos, Theor. Chim. Acta,
1995, 92, 67–81; (c) E. C. Brown and W. T. Borden, Organometallics,
2000, 19, 2208–2214.
8 For discussions of phosphabenzene molecular orbitals: (a) M. H.
Palmer, R. H. Findlay, W. Moyes and A. J. Gaskell, J. Chem. Soc.,
Perkin Trans. 2, 1975, 841–850; (b) G. Frison, A. Sevin, N. Avarvari,
F. Mathey and P. LeFloch, J. Org. Chem., 1999, 64, 5524–5529.
9 M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria,
M. A. Robb, J. R. Cheeseman, V. G. Zakrzewski, J. A. Montgomery,
Jr., R. E. Stratmann, J. C. Burant, S. Dapprich, J. M. Millam,
A. D. Daniels, K. N. Kudin, M. C. Strain, O. Farkas, J. Tomasi,
V. Barone, M. Cossi, R. Cammi, B. Mennucci, C. Pomelli,
C. Adamo, S. Clifford, J. Ochterski, G. A. Petersson, P. Y. Ayala,
Q. Cui, K. Morokuma, D. K. Malick, A. D. Rabuck,
K. Raghavachari, J. B. Foresman, J. Cioslowski, J. V. Ortiz,
J. V. Stefanov, G. Liu, A. Liashenko, P. Piskorz, I. Komaromi,
R. Gomperts, R. L. Martin, D. J. Fox, T. Keith, M. A. Al-Laham
C. Y. Peng, A. Nanayakkara, C. Gonzalez, M. Challacombe,
P. M. W. Gill, B. Johnson, W. Chen, M. W. Wong, J. L. Andres,
C. Gonzalez, M. Head-Gordon, E. S. Replogle and J. A. Pople,
Gaussian 98, Revision A.6; Gaussian, Inc., Pittsburgh, PA, 1998.
An LDA (0.88 mmol) solution was prepared by adding iPr2-
NH to BuLi at Ϫ78 ЊC. After stirring for 3 minutes, the LDA
solution was cooled to Ϫ85 ЊC and (E,E)-1-(prop-2-ynyloxy)-
hexa-2,4-diene31 (100 mg, 0.73 mmol) was slowly added as a
solution in Et2O (1.1 mL). After 3 min of stirring, TMSCl was
added dropwise followed by the dropwise addition of HMPA as
a solution in Et2O. The reaction was allowed to warm slowly
to room temperature with stirring for 30 min during which
time the cloudy solution became a milky white suspension. The
reaction was quenched with 3 M HCl and extracted with ether.
The extracts were washed with saturated NaHCO3 then brine,
dried over MgSO4 and concentrated. The crude product was
purified by flash chromatography (5% Et2O–pentane) to give
5 as a yellow liquid in 89% yield. 1H NMR (200 MHz, CDCl3)
δ 6.3–5.95 (m, 2H), 5.85–5.55 (m, 2H), 4.12 (s, 2H), 4.05 (d,
J = 6.5 Hz, 2H), 1.75 (d, J = 6.72 Hz, 3H), 0.17 (s, 9H).
13C NMR (125 MHz, CDCl3) δ 134.5, 131.1, 130.6, 126.2,
102.0, 91.6, 70.3, 57.9, 18.4, 0.2.
Thermal control reactions
For purposes of comparison a sample of 6 was synthesized
thermally from dieneyne 5 (20 mg, 0.09 mmol) by heating
in benzene (9.6 mL) at 80 ЊC in a resealable pressure tube. In
addition, the thermal background reaction was determined in
the solvent of choice for the phosphabenzene reactions (cyclo-
hexane). Dieneyne 5 (75 mg, 0.36 mmol) was dissolved in
cyclohexane (36 mL) and set to reflux. Monitoring of the reac-
tion progress was performed by 1H NMR which indicated
16% conversion after 20 h with significant decomposition. Con-
version was determined by comparing the relative integration
for the peaks at δ 1.8, d{–CH3(5)} and δ 1.19, d{–CH3(6)}.
Upon chromatography, a 10% isolated yield of 6 was obtained.
J. Chem. Soc., Perkin Trans. 1, 2002, 439–444
443

View Article Online
10 These orbitals were generated using MOLDEN: G. Schaftenaar,
J. H. J. Noordik, Comput.-Aided Mol. Des., 2000, 14, 123–134.
11 C. A. Tolman, Chem. Rev., 1977, 77, 313–348.
12 PM3 geometry-optimized structures were used as calculated with
SPARTAN v5.1.2 (Wavefunction, Inc.; 18401 Von Karman Avenue,
Suite 370; Irvine, CA 92612, USA). Later calculations indicated that
these geometries were almost identical to the HF/6-31G* or B3LYP/
6-31G* geometries.
13 J. Deberitz and H. Nöth, J. Organomet. Chem., 1973, 49, 453–467.
14 For a listing of the CO stretching frequencies for a number of
LؒCr(CO)5 complexes see: W. Strohmeier and F.-J. Müller, Chem.
Ber., 1967, 100, 2812–2821.
provided low efficiency while causing substrate decompostion and
formation of aromatic by-products. PCy3 provided moderate
efficiency but caused the formation of conjugated by-products.
P(O-o-C6H4Ph)3 and P(O-i-C3HF6)3 were both highly efficient,
causing no substrate decomposition or by-product formation.
19 (a) C. Elschenbroich, M. Nowotny, A. Behrendt, W. Massa and
S. Wocadlo, Angew. Chem., Int. Ed. Engl., 1992, 31, 1343–1345;
(b) H. Lehmkuh, P. Rainer and R. Mynott, Liebigs Ann. Chem.,
1981, 1139–1146.
20 G. Märkl, Chem. Unserer Zeit, 1982, 16, 139–148.
21 G. Märkl, F. Lieb and A. Mertz, Angew. Chem., Int. Ed. Engl., 1967,
6, 458–459.
15 Mulliken and natural bond order charges were calculated for the
PM3 geometry-optimized structures using SPARTAN v5.1.2 (see
ref. 12) and for the HF and B3LYP geometry-optimized structures
using Gaussian98 (see ref. 9).
22 The protocol reported for 8 was found to also work with 7 and 9
K. Dimroth, K. H. Wolf, Newer Methods of Preparative Organic
Chemistry, ed. W. Foerst, Academic Press, New York, 1964, Vol. III,
p. 357.
23 Available from Lancaster or prepared following: S. M. M. Elshafie,
Indian J. Chem. Sect. B: Org. Chem. Incl. Med. Chem., 1981, 20,
427–428.
24 S. M. Makin, E. K. Dobretsova, O. A. Shavrygina and L. A.
Kundryutskova, J. Org. Chem. USSR, 1986, 22, 1786–1788.
25 HBF4 was used in place of HClO4 to form the desired salt: A. Baeyer
and J. Piccard, Liebigs Ann. Chem., 1911, 384, 208–224.
26 M. Grayson, J. Am. Chem. Soc., 1963, 85, 175–183.
27 K. Dimroth and W. Mach, Angew. Chem., Int. Ed. Engl., 1968, 7,
460–461.
16 For comparison, B3LYP/6-31G* Mulliken and natural charges,
respectively, were calculated for several other compounds: P(OMe)3,
0.887, 1.603; PPh3, 0.334, 0.902; PEt3, 0.409, 0.845. The NBO values
for this series of sp3-hybridized phosphorus ligands qualitatively
correlate with their χ values (Table 3); however, the charges of the
sp2-hybridized phosphabenzenes do not fall in the range expected
from this series. Comparison of the calculated charges with the
χ values of such divergent structures may not be valid, since the
energy levels and symmetry of the orbitals available for σ-bonding
and π-backbonding will differ substantially between phosphines
and phosphabenzenes and are likely contribute to χ. Within the
phosphabenzene series, the trend seen in the calculated charges does
correlate with the measured χ values.
28 W. C. Still, M. Kahn and A. Mitra, J. Org. Chem., 1978, 43,
2923–2925.
29 Adapted from: H. Jacobi, Liebigs Ann. Chem., 1923, 432, 309–312.
30 A. Markl, F. Lieb and A. Merz, Angew. Chem., Int. Ed. Engl., 1967,
6, 944–945.
17 G. Frison, A. Sevin, N. Avarvari, F. Mathey and P. LeFloch,
J. Org. Chem., 1999, 64, 5524–5529.
18 P. A. Wender and T. E. Jenkins, J. Am. Chem. Soc., 1989, 111,
6432–6434 The results of ligand screening indicated that PPh3
31 Y. I. Nilsson, R. G. P Gatti, P. G. Andersson and J. E. Backvall,
Tetrahedron, 1996, 52, 7511–7523.
444
J. Chem. Soc., Perkin Trans. 1, 2002, 439–444